Benchmark Almanac

ElephantBench

Kimi K3 52.4% leads since August 2026
Not in model cards

Models answer 1,094 closed-book questions about facts with multiple verified but conflicting accounts. A model judge assigns complete, partial, or failed recall.

The benchmark reports recall rates and completeness. Sources and tools are withheld during answering.

Released
August 2026
Kind
Short factual questions
Level
Expert
Scored
Graded by a model
Answers
Short answer
Tasks
All public
Measures
Capability
Language
English
In and out
Text → Text
Interaction
Single turn
Refresh
One fixed set
Run by
Anyone
Strongest model tested
GPT-5.6 Sol #14 on the capability index
Last model added
July 2026
Scores refreshed
November 18th, 2026

In the authors’ words

Factual question answering (QA) typically assumes a single canonical answer, obscuring whether large language models (LLMs) retain divergent accounts of long-tail facts. To address this gap, we introduce ElephantBench, a closed-book knowledge probe comprising 1,094 questions generated through an auditable graph-based pipeline. The pipeline retrieves related documents from a low-exposure web corpus, identifies naturally occurring disagreements, and converts them into multi-account QA records. Each answer is verified against the originating documents and authoritative public web sources and is then reviewed by human annotators. Across 32 models, even the strongest model recovers both accounts on only 52.4% of questions, while on nearly all remaining questions it recalls one account but omits the other. Scaling model size and inference-time reasoning improve recall but do not eliminate this incompleteness. Corpus analysis further shows that exposure imbalance favors the dominant account, whereas greater minority-side exposure is associated with more complete recall. These findings establish ElephantBench as a reproducible knowledge probe for diagnosing epistemic myopia in parametric memory. More broadly, our graph-based benchmark construction pipeline provides an efficient and scalable way to turn long-tail corpora into source-traceable knowledge probes, supporting efforts to evaluate and advance the epistemic rigour of next-generation LLMs. Code is available at https://github.com/Tencent/ElephantBench.

Abstract of Blind Men and the Elephant: Probing the Epistemic Myopia of LLMs under Long-Tail Divergent Knowledge (arXiv:2608.28478, CC0)
32
Entries on its own leaderboard, 29 models we track

Kimi K3 leads with 52.4% among models released by July 2026 Kimi K3 leads with 52.4% among models released by July 2026

Best score per entry from the benchmark's own leaderboard, read on November 18th, 2026. Where a model ran with several settings or harnesses, we keep its best. Names in gray aren't in our model list: fine-tunes, agent systems, products and older models. They don't count toward leaders. This view shows only models we track, each at its best.

The newest model with a score here came out in July 2026, and nobody has tested a newer one. Among the 20 highest-ranked models on the capability index, these have no score yet:

  1. 1 Kimi K3 52.4%
  2. 2 Gemini 3.1 Pro 50.4%
  3. 3 GPT-5.5 50.2%
  4. 4 Gemini 3.5 Flash 48.6%
  5. 5 GPT-5.6 Sol 45%
  6. 6 Claude Opus 4.8 44.1%
  7. 7 Tencent Hy3 42.6%
  8. 8 Nemotron 3 Ultra 38.1%
  9. 9 Claude Sonnet 5 37.8%
  10. 10 GLM-5.2 37.3%
  11. 11 MiniMax-M3 36.2%
  12. 12 Qwen3.5 397B-A17B 32.3%
  13. 13 DeepSeek-V4-Pro 32%
  14. 14 DeepSeek-V4-Flash 30%
  15. 15 Qwen3.5-122B-A10B 28.2%
  16. 16 gpt-oss-120b 22.5%
  17. 17 Qwen3.5-35B-A3B 21%
  18. 18 Qwen3.5-27B 19.5%
  19. 19 Qwen 3.6 35B-A3B 17%
  20. 20 Seed-OSS-36B 15.4%
  21. 21 OLMo-3.1-32B 15%
  22. 22 Gemma 4 26B A4B 13.7%
  23. 23 Qwen3.5-9B 12.3%
  24. 24 Gemma 4 31B IT 11.9%
  25. 25 Qwen3.6 27B 10.2%
  26. 26 Qwen3.5-4B 9.69%
  27. 27 gpt-oss-20b 9.14%
  28. 28 Gemma 4 12B 5.85%
  29. 29 OLMo-3-7B 5.21%
  30. 30 Gemma 4 E4B 2.74%
  31. 31 Qwen3.5-2B 1.65%
  32. 32 Gemma 4 E2B 1.37%

Source: Official ElephantBench leaderboard

#EntryScoreSettingDateSource
1 Kimi K3 52.4% Jul 2026 Official leaderboard
2 Gemini 3.1 Pro 50.4% Feb 2026 Official leaderboard
3 GPT-5.5 50.2% Apr 2026 Official leaderboard
4 Gemini 3.5 Flash 48.6% May 2026 Official leaderboard
5 GPT-5.6 Sol 45% Jul 2026 Official leaderboard
6 Claude Opus 4.8 44.1% May 2026 Official leaderboard
7 Tencent Hy3 42.6% Jul 2026 Official leaderboard
8 Nemotron 3 Ultra 38.1% Jun 2026 Official leaderboard
9 Claude Sonnet 5 37.8% Jun 2026 Official leaderboard
10 GLM-5.2 37.3% Jun 2026 Official leaderboard
11 MiniMax-M3 36.2% Jun 2026 Official leaderboard
12 Qwen3.5 397B-A17B 32.3% Feb 2026 Official leaderboard
13 DeepSeek-V4-Pro 32% Apr 2026 Official leaderboard
14 DeepSeek-V4-Flash 30% Apr 2026 Official leaderboard
15 Qwen3.5-122B-A10B 28.2% Feb 2026 Official leaderboard
16 gpt-oss-120b 22.5% Aug 2025 Official leaderboard
17 Qwen3.5-35B-A3B 21% Feb 2026 Official leaderboard
18 Qwen3.5-27B 19.5% Feb 2026 Official leaderboard
19 Qwen 3.6 35B-A3B 17% Apr 2026 Official leaderboard
20 Seed-OSS-36B 15.4% Official leaderboard
21 OLMo-3.1-32B 15% Official leaderboard
22 Gemma 4 26B A4B 13.7% Apr 2026 Official leaderboard
23 Qwen3.5-9B 12.3% Feb 2026 Official leaderboard
24 Gemma 4 31B IT 11.9% Apr 2026 Official leaderboard
25 Qwen3.6 27B 10.2% Apr 2026 Official leaderboard
26 Qwen3.5-4B 9.69% Feb 2026 Official leaderboard
27 gpt-oss-20b 9.14% Aug 2025 Official leaderboard
28 Gemma 4 12B 5.85% Jun 2026 Official leaderboard
29 OLMo-3-7B 5.21% Official leaderboard
30 Gemma 4 E4B 2.74% Mar 2026 Official leaderboard
31 Qwen3.5-2B 1.65% Feb 2026 Official leaderboard
32 Gemma 4 E2B 1.37% Mar 2026 Official leaderboard
#ModelScoreSettingDateSource
1 Kimi K3 52.4% Jul 2026 Official leaderboard
2 Gemini 3.1 Pro 50.4% Feb 2026 Official leaderboard
3 GPT-5.5 50.2% Apr 2026 Official leaderboard
4 Gemini 3.5 Flash 48.6% May 2026 Official leaderboard
5 GPT-5.6 Sol 45% Jul 2026 Official leaderboard
6 Claude Opus 4.8 44.1% May 2026 Official leaderboard
7 Tencent Hy3 42.6% Jul 2026 Official leaderboard
8 Nemotron 3 Ultra 38.1% Jun 2026 Official leaderboard
9 Claude Sonnet 5 37.8% Jun 2026 Official leaderboard
10 GLM-5.2 37.3% Jun 2026 Official leaderboard
11 MiniMax-M3 36.2% Jun 2026 Official leaderboard
12 Qwen3.5 397B-A17B 32.3% Feb 2026 Official leaderboard
13 DeepSeek-V4-Pro 32% Apr 2026 Official leaderboard
14 DeepSeek-V4-Flash 30% Apr 2026 Official leaderboard
15 Qwen3.5-122B-A10B 28.2% Feb 2026 Official leaderboard
16 gpt-oss-120b 22.5% Aug 2025 Official leaderboard
17 Qwen3.5-35B-A3B 21% Feb 2026 Official leaderboard
18 Qwen3.5-27B 19.5% Feb 2026 Official leaderboard
19 Qwen 3.6 35B-A3B 17% Apr 2026 Official leaderboard
20 Gemma 4 26B A4B 13.7% Apr 2026 Official leaderboard
21 Qwen3.5-9B 12.3% Feb 2026 Official leaderboard
22 Gemma 4 31B IT 11.9% Apr 2026 Official leaderboard
23 Qwen3.6 27B 10.2% Apr 2026 Official leaderboard
24 Qwen3.5-4B 9.69% Feb 2026 Official leaderboard
25 gpt-oss-20b 9.14% Aug 2025 Official leaderboard
26 Gemma 4 12B 5.85% Jun 2026 Official leaderboard
27 Gemma 4 E4B 2.74% Mar 2026 Official leaderboard
28 Qwen3.5-2B 1.65% Feb 2026 Official leaderboard
29 Gemma 4 E2B 1.37% Mar 2026 Official leaderboard

Number one over time

Lab claims count from the day the card came out, independent runs from the model’s release.

Moonshot AI
Own leaderboard Not tested since Aug 2026
Kimi K3 52.4%

Source: Official ElephantBench leaderboard

Scores over time

All the scores we have, one dot each, so a model tested at several settings appears more than once.

Plot with other benchmarks
On the benchmark’s own leaderboard Best so far
0% 20% 40% 60% 80% 100% 2026 Near the ceiling (85%) Kimi K3 Gemini 3.1 Pro Qwen3.5 397B-A17B gpt-oss-120b 0% 20% 40% 60% 80% 100% 2026 Near the ceiling (85%) Kimi K3 Qwen3.5 397B-A17B

Source: Official ElephantBench leaderboard

No lab model card reports it yet

We read the headline results of 540 launch posts from 57 labs, and none of them list this benchmark.

More in Factuality

All 164 in Factuality
  • SimpleVQA In use Models answer 2,025 short factual questions about images across 9 tasks and 9 topics. An LLM judge grades answers against fixed references. Qwen3.7-Plus 81.7%
  • HallusionBench In use Models answer 1,129 yes or no questions about 346 images, built in pairs to expose language hallucination and visual illusion. Scored by accuracy. Keye-VL-2.0-30B-A3B 73.5%
  • SimpleQA Verified In use 1,000 human-verified, de-duplicated short-answer factual questions measuring parametric knowledge. Gemini 3.1 Pro 77.5%
  • AA-Omniscience v1.1 In use Models answer 3,000 knowledge questions in six domains without tools and may abstain. The index adds correct answers and subtracts wrong ones. Inkling 2 Small 16.1
  • GaRAGe In use Models write long-form RAG answers to 2,366 questions with human-annotated grounding passages. Scoring checks whether answers use the relevant passages. Amazon Nova Pro 60.7%
  • Kimi Hallu Bench In use Models answer hard prompts drawn from practical work tasks. Fact-checking agents score the share of claims that are hallucinated. GPT-6 Astra 4.20%

Data audit

All 10 checks pass. Every benchmark

Catalog record 2 pass
  • Pass Task access All tasks are public.
  • Pass Confidence Checked against a primary source.
Scores 4 pass 1 note
  • Pass Independent runs Scores from the official leaderboard (32 entries).
  • Pass Models scored 29 models we track, plus 3 other entries on its leaderboard.
  • Pass Newest score The newest is from July 2026, for Kimi K3.
  • Note Frontier models None of the top 10 has a score. The strongest tested is GPT-5.6 Sol (#14).
  • Pass Headroom The best is 52.4%, under 85%, so it still separates models.
Leaderboard import 3 pass
  • Pass Importer We read the board on tencent.github.io: 29 rows for models we track and 3 other entries at the last read.
  • Pass Last read November 18th, 2026.
  • Pass Read worked The last read worked.
Consistency 1 pass 1 note
  • Note Lab claims No model has both a lab number and an independent run.
  • Pass Known issues No published audit found problems.

How we check